The Geometry of Prompting: Unveiling Distinct Mechanisms of Task Adaptation in Language Models

TL;DR

This paper explores how different prompting methods influence the internal representations of decoder-only language models, revealing distinct mechanisms of task adaptation and emphasizing the importance of input and label semantics.

Contribution

It introduces a statistical physics framework to analyze the geometry of representations, uncovering diverse mechanisms behind prompt-induced task adaptation in language models.

Findings

Different prompting techniques operate through distinct representational mechanisms.

Input distribution samples and label semantics are crucial in few-shot learning.

Interactions between tasks can be synergistic or interfering at the representational level.

Abstract

Decoder-only language models have the ability to dynamically switch between various computational tasks based on input prompts. Despite many successful applications of prompting, there is very limited understanding of the internal mechanism behind such flexibility. In this work, we investigate how different prompting methods affect the geometry of representations in these models. Employing a framework grounded in statistical physics, we reveal that various prompting techniques, while achieving similar performance, operate through distinct representational mechanisms for task adaptation. Our analysis highlights the critical role of input distribution samples and label semantics in few-shot in-context learning. We also demonstrate evidence of synergistic and interfering interactions between different tasks on the representational level. Our work contributes to the theoretical understanding of large language models and lays the groundwork for developing more effective, representation-aware prompting strategies.